This is your laptop... on AI

And build your system. Sonos. Love the sound. Feel the system. Computers are getting smarter, Silicon Valley's getting weirder, and AI is taking over. At least that's what everyone keeps telling us. Welcome to Machine Gods, a new NPR podcast that will be your field guide to the AI-powered future. We're veteran tech journalists with front row seats to the industry. And twice a week, we'll make sure you get smarter about AI, so you know what's real, what's hype, and what might come next. Listen to Machine Gods from NPR on the NPR app, or wherever you get your podcasts.

Support for the show comes from ServiceNow. AI is moving fast across the enterprise, but without visibility, it's just chaos. Different tools, different models, different teams using AI in completely different ways. ServiceNow turns that chaos into control. With the AI control tower, you see all your AI across the business in one place. What it's doing, what it's done, and what it's about to do. So you stay in control. To put AI to work for people, visit servicenow.com.

Hello and welcome to the Verge Cast, flagship podcast of the Fox Media Podcast Network. Still, for now, who's to say? I'm your friend David Pierce. Neal Abtels here. Hey, buddy. Hello. You are, for those listening, I would say you appear to be in the cabin of a very large yacht. It's just the vibe happening behind you. Where are you in the world right now, Neal? Fully the vibe of this house. So I'm married to a literal farmer's daughter. Becky has farms on both sides of her family here in Illinois. I'm in Illinois, in the middle of nowhere, Illinois. And so we're on the farm on her mom's side, which her uncles have turned into like a beautiful house.

Like, the original house is there, but then they just added crazy house next to it. And so I'm in the basement office of her uncle's house on the farm. But, like, her whole family is here. There's just kids running around everywhere. And then her uncle's partner has a parrot that he got in 1989. Whoa. This parrot owns the house, right? The parrot is, like, a longer patriarch of this family than anyone else at this point. And so Norton is just screaming whenever he sees children. The children have all learned that they can sing at Norton. Sorry, Norton the parrot? He's right. He rules. I immediately know everything I need to know about this parrot by the fact that his name is Norton, and he was born in 1989. Yeah, he's right. It's awesome.

So yeah, we're at this sick house on the farm for the family reunion this week. And so if you hear children, grand pianos, or parrots, just know what's going on here. The big three, as we call them. The farmer trio. That's really good. All right, we got a lot of stuff to talk about. We got a bunch of AI news. We got a bunch of developer conference stuff going on. We got Brendan Carr doing stuff. Brendan is particularly, there might be some straight F-bombs when it comes to Brendan this week. And not for any reasons you might think, just because it's so dumb. It's so, so dumb. People always ask, why is the word dummy? And it's like, sometimes there are weeks where it's just, he's a dummy. This is just what he's doing. Just transparently stupid. Anyway. Yeah.

I'm excited about it. The one bit of business we should get into at the very top is we have new Verge merch. I am the first, I believe, person on earth to be wearing our new flagship podcast shirt, which I'm very excited about. We also have new decoder merch, new version history stuff. We have shirts, we have mugs. We have lots of cool stuff coming, but a thing people have been asking for for a long time is like show merch that's cool and good and high quality. So we have a bunch of that on the Verge shop. Go get it. I have decoder merch before Neelai does, which makes me feel really cool and powerful. I'm very excited about it. I did not realize that we were going to make a people do not yearn for automation t-shirt, which I'm very excited about.

It's great. I have it over there. There's also, I will warn you, going to be a costume change before Brennan cards a dummy today. Just as a brief teaser. Can I show you my merch, which we're not selling, but I just realized. Here's the coffee cup I pulled out of the... It was the biggest one in the cabinet. It's perfect. Farm Nilai is like a whole different vibe. I'm never coming back.

It's good stuff. All right. We should talk about AI stuff because there's a lot of AI stuff going on this week. I think we're like deep in developer conference season. And I would say the theme of developer conference season so far between Google I.O. This week was Microsoft Build and next week is WWDC is basically just what do people want to do with AI? Yeah. There are lots of ideas about business apps and how you can sort of interact with all of the tools that your company uses. Then there are questions of like, what do people want these tools to do? There is this increasing sense that actually this technology is very good at a lot of things. It has really developed in huge ways over the last 12 months. Do you want this?

over and over the question from all these companies. Here's a product we made. Do you like this? Is this the one? Yeah. And the one I want to start with is the one that a couple of folks on our team have been testing, which is Gemini Spark. This is like the idea right now, I would say, right? Is how do we take a bunch of the information that exists about you around our services, whether it's like Google Workspace stuff, if you're Google or it's Microsoft Office, if you're Microsoft, and give you ways to like interact with, but also just have an AI agent go use those tools on your behalf. Like Spark's idea is to make calendar events for you and write emails for you and make Google Docs for you.

which Microsoft launched this week is exactly trying to be the same thing for all of Microsoft's apps. My first question for you, Eli, is like, does this idea make sense to you? Like AI agent for your workspace. Sure. There are three pretty obvious eras of AI that we've passed through. There was the first one, the sort of chat GPT moment. You have a chat bot. It will do everything, including be your boyfriend. Shout out Kevin Roos. You'll be your boyfriend in the pages of the New York Times. That was one era, right? We're going to kill Google search. Google search is the best business in the history of the world. And a bunch of companies are going to show up and try to do that. The second era was reasoning models, right? That was last year, maybe a little bit before that. We're going to do long, long chains of thought. I'm putting reasoning and thought in quotes if you're listening to this audio. I hope you can hear the disdain dripping off my voice as I talk about these words that have been co-opted.

you know, chains of thought means that model is basically going to prompt itself over and over again, and it's going to talk to itself a lot before it issues you an answer in the form of some research talk. And this got everything a lot farther, right? You could see the models got more powerful. They ended up in more enterprise use cases. And I would say the consumer use for reasoning models just never appeared. You could pay Google or OpenAI or whoever, and I used a reasoning model to negotiate a car lease. This is not a thing that normal people should do.

And then now we're at agents where, okay, open claw happened. You can put a reasoning model next to your actual computer and your actual applications on your Mac mini or your MacBook that's cracked open while you walk through the airport. And it's going to actually use those applications. And you need pretty fast CPU to run the model. And then you're going to send all the inference up to the cloud and the models are on the cloud. And this is neat. It's just the next turn. It's the next era. And so every one of these developer conferences has been about agentic AI, which is just the next turn of letting the model do more stuff.

like run longer, burn more tokens, all that. And then I think we've talked about this a lot. Somewhere embedded in there is, oh, we can write code now. We can write code well enough, not great as we will discuss in some other contexts, particularly security, but we can write code now, which means not only can we use the computer, we can write applications for the computer to run that we can then use ourselves. And this is, you know, this is why Demis Isadis is talking about the foothills of the singularity, right? Like now the computer is not only using itself, but it's programming itself. And if you have software brain, the computer is now alive, like fully and completely alive. It can tell you it loves you and it can write code.

What else is there? That's what makes us human, right? Those are the two things that make us human, apparently. So everyone's just running at it. I think Google's advantage is all of their software already lives on the web, in the cloud. And your data is already there, as you have discovered in your testing. Like Gmail is just a cloud application that already runs on Google servers. They don't need a Mac mini. You're just like, Gemini Spark, go do computer stuff. And it's like, I already do that. I think Microsoft would like this. I think all of these AI companies have built browsers are trying to get there where they can use a browser on your behalf. But Google just has so much of it already that they don't need to overprove the concept. They're like, yeah, we can just use Gmail for you.

We already do that. And it seems like that is incredibly effective. I was blown away by how well it works. Like I, the experience I had with Spark was basically try to do a number of the things that it tells you you can do, right? Where you say like, oh, go make, one of the things that Josh Woodward who runs this team said at IO was like, go make all of my meetings with Sundar Pichai pink so that they like pop on my calendar. Relatively easy, straightforward AI task, right? And that's the kind of thing Spark does very well. It's the like little tiny bits of busy work that.

these tools are supposed to do for you all the time. And like you said, all this stuff exists within Google already. So putting the pieces together is actually a bunch of people at Google having to have meetings more than it is like systemic technological breakthroughs. They've solved this problem in some really interesting ways. The thing that most sort of messed with my brain as a person was taking it outside of the work context. And like, this might have been the story I wrote is I fed it basically a trip planning scenario, right? Like we've talked about this on the show. All anybody can think about is what if AI could book you plane tickets, right? Like that's everyone's AI idea. And I just did the thing that everybody always does in these demos, which is like, okay, I'm going to Hershey, Pennsylvania.

the weekend of July 18th. Can you just build me a whole itinerary based on just that level of information? I was like, I'm going with my wife, my two kids and my dog. Can you build us an entire itinerary? And it went, thought about it for a few minutes. And ordinarily what comes back in that case is like, here are the top eight things on TripAdvisor to do in Hershey, Pennsylvania, right? It's just sort of a very basic, like read one travel guide answer.

this was very different it went through and figured out like okay it understood that um i had concert tickets for saturday night which is a thing that exists in my email and not anywhere else in the interactions i've had with google so it went and pulled that and was like oh this it figured out this is why we're going to hershey because we have tickets for saturday night uh it knows a series of things that i was not aware that google knew like the names and ages of both of my children um the dietary restrictions of my wife which is one that i sincerely still cannot figure out like i've never emailed with anna about her dietary restrictions i don't i don't i don't know um I just don't, I don't know. And it went through and found restaurants that have all the things that our kids need. It went and found the pet fees for various hotels for things that we needed. It went and found how much Hershey Park would cost based on the ages of my children. Like it came back with this incredibly detailed, personalized, really, really, really useful itinerary that like frankly scared the shit out of me. Like in a way I was totally unprepared for this thing that was like really, really helpful and personalized and good. And you would have gotten something very different, right? Like if you had asked it the same question about the same days, Spark would have given you a completely different set of things to do.

which on the one hand is completely correct, right? Like this is a huge advancement in this thing that these tools are trying to do in the fact that it can do that. The fact that it's not just reading me the top of TripAdvisor, but it's actually giving me, David Pierce, real recommendations based on my life and situation. So powerful. And I got to the end of this and I was like, I need to throw my computer into the ocean, delete all of my Google information and never, ever, ever go online ever again.

And I just came to the conclusion that this is the thing that we all have to decide whether we're comfortable with, because this is the trade. That thing worked astoundingly well, and it only worked astoundingly well because it is a terrifying privacy hellscape in which Google knows every piece of information about me, regardless of whether I've actually proactively told Google that information. And I still, I've been sitting here for a week looking at this itinerary, not knowing how to feel about it. All right, let me make the bull case and the bear case. So the bear case for this, the negative case for this, is this is exactly how people feel about advertising online that makes them think.

their phones are recording them all the time. It is exactly that feeling and just magnified to its highest level, right? The thing is listening to you and Anna talk about meal planning and it's like, I have deduced your dietary restrictions. It's been spying on you the entire time. And the flip side of that, the pro case is, well, all we got out of all that surveillance in the past was you glanced at Wayfair one time in your life and this chair is going to follow you until you die. And that has never been good enough. And maybe this is the thing that's good enough, right? This is the thing that makes it all worth it. I don't think I'm at, this is the thing that makes all the surveillance worth it.

But that's the case, right? There's finally a product that... makes you useful items out of all the surveillance, not just advertising. Right. Now, is Google going to lace a bunch of advertising into your itinerary? Yes, they sure as hell are. Of course. Yeah. And like, maybe that will break, but you know, we're in the like honeymoon period where in particular Google is very, very good at losing money until it's time to make money. Yeah. And so the product will be great until it's not great. And maybe we should just enjoy the honeymoon period. But that's the, those are the two sides of the coin, right? Is it is exactly the feeling of, oh, my phone has been listening to me, which everyone hates. And, but maybe this time the payoff is worth it. I don't, by the way, I do not think the payoff is worth it.

I was just going to say, yes, I agree with all of that. And I think coming out of this 15-year run where Mark Zuckerberg and Sundar Pichai and everybody else in this space have been trying desperately to convince us that what we want is targeted ads, that actually our experience of the internet is better if the ads are targeted. And I don't think people agree with that, generally speaking. I've been thinking about a thing Sundar said to me in that Decoder conversation since it happened. it's the same conversation you and I have been having, whether the numeric indicators of success are enough, right? Whether this concept of revealed preferences through click data actually tells you about how anyone feels about anything. And so, you know, I handed Sundar my phone and was like, look at this search result for best Chromebook with me. And in the moment, I actually didn't think about like his initial response, right? I handed him the phone and he was like, great, we're doing this again. And we like got into it. And I went and watched it back and he was saying something as I was handing him the phone that I'd missed in real time. He's like, we actually have a very scientific and rigorous way of measuring this, right? Like the thing you and I are about to do has nothing to do with Google's process for this. There's some...

method inside of Google where they actually measure the results page and then they determine if it's good. And they have trillions of queries and they do this. And I'm sure at this point, Gemini is like all in there being like, this one's good, right? And the gulf between Google's quantitative evaluation of whether the search page or the search experience is good, which they are absolutely relying on and maybe totally dependent on. And our qualitative experience of those products is wider than ever. Because every one of these platform companies will tell you without a shadow of a doubt that people love targeted advertising. That they keep making more money, that the advertisers keep dumping money into those systems.

They keep sending customers to small businesses. And then everyone else is like, yeah, but my phone's listening to me. And you will never convince me otherwise, Adam Mosseri. Adam Mosseri goes on his personal Instagram like once a month. You're like, I'm not listening to you. No one believes him. It's the same with Google, right? They're totally reliant on their quantitative metrics. All the numbers prove that you love this. And then everyone, I think, is sitting there. I mean, like, I actually don't love this. Like, maybe if you filter me into the aggregate of your billion users, you can say that we love this. But I don't love this.

And I count. And there's something there that all of these companies at their scale are going to have to start to reckon with, I think, as they start using more and more of the data to construct these personalized experiences. Because there's no way to make me or you or someone a personalized Spark experience and then say, the data says you love this. Right. Because they made that for you.

And something in there is going to break. Like you can feel it, right? Like this page statistically is the best page for everyone is a very different argument than this page is the best page for you. Because statistically, all the other pages we've made for other people are the ones they like. I can't even like say the sentence. It makes no sense. Well, right. And because it's impossible to do in a qualitative way, right? Like Google doesn't actually know or care if you like it. And the LLMs don't know or care about anything. Right, right. And I think like I've spent so much time recently trying to think about how to.

answer the question of like why why do people feel differently about technology than they used to and it's like it's big and messy and mealy in my head but like i think one reason there has been this giant pushback against data centers is it is one specific object that people can be mad about and i think so much of this ai stuff is messy and confusing and it's hard to it's hard to peel parts of it away from each other and it's hard to opt out and it's hard to get away from and it's like it it is easy to be like what i don't want is this gigantic building that is offering me something i don't want in exchange for literally the resources of my community like that it is a it is a physical instantiation of this thing that is otherwise really hard for people to like get their arms around. And so I think I've started to look at people's opinions about data centers very straightforwardly as actually their opinions about like Google's AI overviews, right? Which is just like, this isn't good enough. I don't want this thing in my life. I don't want it in my community. This thing you're saying will take all the jobs away when I use it at my job is not good. Yeah, and it makes me feel bad when I use it because everything that I hear suggests that it's making me dumber. That's a great point, David. You were right to point that out.

That's not just right. It's terrific. Yeah, but I keep coming back to the idea, again, like this Spark experience I had was so good. And I think the reaction to the story that I wrote has been really interesting because it hits these lines exactly, right? Where it's like, on the one hand, there is a sense of like, no, I don't want this. This is a success story that sucks and I hate it. And I think then there was another version of this that was like, yeah, this is exactly the thing that we've been trying to do. This is the whole promise of AI spelled out for you. You didn't have to go do eight hours of...

really irritating Airbnb research because it found you the Airbnb that you should book. Like this is the moment that we're coming to with AI in a really interesting way, which is the tech is starting to work, right? Like it just is. We are at a moment now where all of the people who were like, it's just fancy autocorrect. It will never be the thing. Like I hate to be the bearer of bad news on this, but like the tech kind of works now. It's not digital Jesus. It's not going to remove everyone's jobs in a week, but like it can do things now. It really can. I've seen it. I use these tools all day, every day. They work. And we are now confronted with the question in this like hugely existential way of do we want this?

Like, do we want this? And we have to ask it now. And we have to do things about it now. And overwhelmingly, the answer is no. And I think we're going to get to a point where it's like, this stuff works and I still don't want it. And that, whether or not that message ends up being more powerful, I think is going to be sort of the question about AI for me for the next little while. Like, it sucks and I don't want it. It's whatever. I'm going to caveat works with your own feelings about AI that you have expressed on the show many times. It works for business. And it works when you treat your own life like business, right? Consumer applications work if you compress yourself into being a travel agent.

Do you know what I mean? Sure. Like, that's great. And this is why I'm always yelling about software brain. It works if you conform to how it wants you to operate, which is very easy for corporations to do, right? Any corporation is like, oh, there's a 20% efficiency increase. We will immediately change the process of the corporation to go get that change. And that's the entire enterprise software market works in this way. Normal people just don't, right? And you say data centers are a reflection of how people feel about AI overuse, but I also fully disagreed with that when I put that to him. I think the opposition data centers is how all of the products make all the people feel, which is stupid.

All the time. Interesting. Right? Even though they're like synchophantic and they try to make you feel like a genius, at some point you hit the brittleness, particularly free models, or you hit the like, this isn't the answer I wanted, but it's been presented in such a confident form that I'm just going to walk away from you. Right. Which is something like what Gemini Spark is doing. Right? It made you a beautiful plan. And to argue with that plan, like, it's not like talking to your actual travel agent.

It's not like talking to your friend who helps you plan a trip because they're really good at credit card points or whatever. It's like, I have to prod the computer in some way to not screw up the first half of the trip because I want to change a restaurant in the second half of the trip. And like, now you're a computer programmer. You might be using natural language, but you have to conform yourself to how Gemini wants you to issue these instructions. And that stuff isn't good enough yet. It is great in the business context because you have every incentive to be good at your job and use these tools in that way. And I'm just pounding the table. Making people's lives feel like work makes for horrible consumer products. These consumer products are not anywhere close to good enough because...

All of them are in the framework of what if more of your life looked like using the computer at work? And I just think that's a mistake. Even the personal contact stuff that Google's doing, that's a straight lift from how enterprises think about using AI because they have these vast stores of data and all of these databases or their data lakes or all this enterprise IT vocabulary. None of that data talks to each other. All of their middleware products that are supposed to connect it all are expensive and don't work. And now you can just let loose Gemini for work on your vast interconnected stores of databases and it will start generating business insights at you. And all of this seems very exciting. The leadership conferences are more energized than ever about this stuff. And no human being in their lives is like, boy, my data is in several unconnected databases.

If only Gemini could read them all and deliver action ones. Like, what? No, man. I just want to drink a bunch of Miller Lights and chill out. Like, there's something there that's missing forever, you know? Yes, I think that's right. But also, I think the belief that these companies, increasingly, is that you're wrong. And that actually, the ability to one-shot me a useful, personalized trip planner is going to cross that chasm. I just want to read you a quote from, so I interviewed this woman, Gazde Ozner, who's a product manager at Google. We were talking in the context of this app called Dream Beans that Google came out with this week. I don't know if you saw this. The idea is, it's essentially a...

proactive list of, they call them stories, of AI-generated things every day. I encourage you to go look at the thing that I posted on the site, which is one of the things it does is it generates pictures of you for a lot of these stories, which is just extremely rude about the size of my forehead in most of them. But the idea is like, okay, how do we put all of this data together for you in sort of a proactive, helpful, useful way? So sometimes it's like, here's a bit of information you should know for today, or like a fun way to tell me to update all my computers ahead of WWDC. It's literally one of the things it gave me. And she said this thing.

She said she was originally on the personal intelligence team at Google. And she said, we realized that each service holds a specific piece of your life. Gmail has mostly your logistics and receipts. Photos has your visual memories and the people you care about. YouTube has your interests and search reflects what you're curious about. So there's definitely something interesting in connecting all of that. Software brain. That is the most software brain, software brain of all time. You are just a bunch of different databases that Google runs. I'm not positive she's wrong. Like, I think I'm positive she's wrong. You think YouTube is a complete list of your interests? I think it's closer than just about anything else you'd find. No way, dude. It's just like, no, that's absolutely not true.

Does YouTube know that you just reviewed a whole bunch of phones and went back to iOS or Android? Probably. How would it know this? based on a bunch of things that I did and the Apple store receipt in my Gmail. I'm telling you, it's sitting there, but it doesn't matter. Google has all of them, right? And this is the bet. And this is like, again, I bring this up in the context of Microsoft Build where their theory is you can do this at work. And I think fairly straightforwardly, you can do this at work, right? That's what I'm saying, in work, this all makes sense to me. Work is just database manipulation for most people. Right, and you put all of these in structure places, like Notion is deep down this road, right? Like put all of your business stuff into Notion and our AI will just do everybody's job for them. Like that's plausible because everybody is incentivized to make these things in these structured ways inside of their companies.

the most alarming thing about this Spark experience for me was like, it's, it kind of works in my personal life. Like it didn't, I'm going to follow that itinerary pretty exactly because it was right about most things. And that's like, that's one example of many things, but it's like, but then the fact that it worked and it still made me feel bad is just this thing I can't get past. And so I think like for me, it's like, okay, this isn't going to work because everybody's software brain. And I think, I think we are inching to the point where maybe at least Google, maybe not anybody else, but maybe at least Google actually does have enough of this data about me that if it can put it all together in a way that.

feels as useful as this one Spark experience did. And again, this is one experience. Life is full of experiences. This is one. But it did the thing in a way that I had just not experienced before. I can just think of ways that this can break you pretty easily. One, if you don't use Google for everything. I will stick on the...

the idea that YouTube doesn't actually know my interests, but Google needs to pretend it does. Sure. Do you know what I mean? Like, that's very important. Aren't you one of those people who like turned off history on YouTube? Yeah. So you don't get recommendations anymore? I open YouTube and it's basically a dark gray search box and a prompt. It's like, please turn on your history. And I'm like, I will never turn on my history. No, it's absolutely true that all of this only works if you willingly have given your, but also like- If you give your life to Google. So like lots of people don't do that. I made the Gmail decision in 2004.

Like, this is not, people are not making this decision today. People made these decisions a long time ago. Right, but some of those people made the decision to use Yahoo in 2002, right? Like, there's a lot of that in the world, right? Where Google can't make the promise for everybody. And there's the software brain piece of this is they have to flatten you into the databases they have. So I'm just going to keep pushing on the YouTube one. To just flatly say, like, YouTube is where your interests are, is required to make the product work. But it has no reflection on where your interests actually are. Because a lot of your interests are also on Instagram. And a lot of your interests are also on TikTok. And a lot of interests are never expressed to your phone at all. Like, at all.

You know? And like, we're both watching Widow's Bay, right? I don't know that YouTube has any indication that I'm watching Widow's Bay on a totally different streaming service. It does because you say your name at the beginning of the podcast and it's a transcript on YouTube. But that's us. That's us. It's there, dude. Like, this is what I'm saying. But like, most people are not opening their podcast. Statistically, everybody has a podcast. So that's not correct. Like, I'm confident that Google knows I play Madden. Like, it's fine. But that's us, not everyone else. I'm just saying that is the software brain. It's the promise. It is also the peril, right? Where you're like, okay, you're just a bunch of databases. What databases do I have that reflect you? YouTube is your interest. And you are sort of just required to say it.

There's also like, you know, Microsoft had built this week. There's their approach to it, which is to say out loud, we're just focused on the enterprise, right? And we have our new models and we're going to build a bunch of enterprise applications. And because we have so much Microsoft 365 and Azure data about your workplace, we can put out whole new products, concept products of what a computer looks like now that have nothing to do with Windows that are just like agent control panels that you might use to do work. This makes sense, again, I think in the enterprise, right? If you sit down at your corporate job, where like mostly you use a laptop to use software, it does make a bunch of sense to say, oh, I can talk to my agent to go use the software in the databases for me. And I think Microsoft sees that opportunity in like gigantic, interesting ways. But everyone else, this is the rabbit R1. Everyone else is like, we're going to run around and use your apps and services for you as a consumer. And that turns your life into work in exactly the same way. And I just, I don't think people want their regular lives to feel like what they do with their jobs.

Maybe some people do. I know people who love to use Airtable to plan their weddings or whatever. There's reasons for lists and schemes to exist everywhere. But no one wants to be like, how do I get 20% more efficiency in my automated process at home? Absolutely not.

Well, see, I agree. I think this is the this is the turning point. Like these things are going to ship to consumers and it's going to be extraordinarily apparent that a lot of us are reflected in a lot of databases that can be linked together to do this kind of stuff. And then everyone is going to have to choose. Yeah. And there's going to be this question of like knowing everything we know about where this technology came from, who stole what to make it and what it's doing to the environment and to our brains. Do I want really great trip planning service? Yeah. Boy, did this all feel very existential just trying to get like a weekend trip to Hershey, Pennsylvania. But actually, we should we should take a break. And then I want to come back and talk about.

the computer thing you just said because I think that's sort of the other AI story of the week is basically Jensen Wong's theory of the computing future and we should litigate this. So let's take a quick break and then we're going to come back and we're going to talk about computers. Be right back. Support for the show comes from Blinds.com. Sometimes you don't realize how important quality window coverings are until you do something about it. All of a sudden, that blinding light you were dealing with every morning goes away and you think, why didn't I get those blinds replaced sooner? With Blinds.com, you can go full DIY if you want or bring in licensed and vetted pros to install the exact lines your room needs. And right now, you can get $50 off when you spend $500 or more. They have a huge variety of style options at prices that fit any budget. And you can get samples shipped to your door fast and for free. If you need help picking the right style, just book a free consultation with one of Blinds.com's award-winning design experts. That means no awkward home visits.

Just good advice and no pressure. Plus, everything is backed by Blinds.com's sure-fit guarantee. Measured wrong? They'll make it. No returns, no hassle. Visit Blinds.com right now for an exclusive $50 off when you spend $500 or more with code VERGECAST. That's Blinds.com with code VERGECAST.

Support for the show comes from Quince. We all have our favorite clothes we love to break out when things get a little chillier. Maybe it's that cool bomber jacket or the sweater that you could wear literally every day. These are exactly the kinds of pieces that Quince specializes in. Quince offers high quality wardrobe essentials made with premium materials like 100% Mongolian cashmere, organic cotton, and merino wool. So it's been about a year now since I started picking up Quince. And one of the things that they talk about is how they just offer those staples that end up being the things that you go to and grab in your closet time and time again. And honestly, it's kind of true.

For me, that's become the Mongolian cashmere quarters of sweater I got towards the end of last fall. Now that these cooler nights are starting to come back, I'm just genuinely excited to reach into my closet and grab what became my old reliable for last year. Find your next fall favorites at Quince. Download the Quince app for app-exclusive offers or go to quince.com slash verge. Get free shipping on your order and 365 day returns. Now available in Canada and the UK too. That's Q-U-I-N-C-E dot com slash verge.

Support for the show comes from Bitwarden. October is Cybersecurity Awareness Month, and here's an upgrade that actually gives you something back, time. With fewer password resets, less typing thanks to autofill, and no more trying to remember which variation of your dog's name you use this time. Bitwarden is a trusted open source solution for passkey, password, and secrets management. It makes secure password management available to everyone with a free forever plan for individuals. That way, all users have a reliable way to create, manage, and autofill strong and unique passwords across accounts and devices.

Bitwarden uses end-to-end encryption to protect vault data, while secure autofill helps reduce phishing risk by matching credentials to the right sites, giving users greater peace of mind when logging in. Bitwarden is trusted by over 15 million users and 80,000 businesses across 180 countries and 50-plus languages. Get started with a basic free plan for individuals or protect your business today with a free trial. This Cybersecurity Awareness Month, check out Bitwarden. Visit bitwarden.com to get started.

All right, we're back. So, Nila, you kind of alluded to this, but the thing I want to talk about here is Jensen Wong's theory about computers. In addition to being Microsoft Build this week, it was also Computex, and NVIDIA announced this chip called the RTX Spark that it thinks is the new consumer chip from NVIDIA. They're going after Apple Silicon, they're going after Intel, they're going after AMD. NVIDIA's really feeling itself here and just trying to take over the computing industry. But in the course of all of this, Jensen Wong was up on stage and he had a long digression into what he believes to be the future of computers. Can I just play you this clip? It's long.

but it is the most Vergecast-y thing anyone has said on stage in a very long time. Can I just play this for you? Yes, let's see it. Okay. About three years ago, I talked to Satya and I said, in the future, AI wants to run on a device too because we want to have assistants running with us all the time. Right now, well, for you, all of your laptops are here. My laptop is in the room. If I want to talk to my laptop today, I got to wait until I get back to my room. In the future, if I need my laptop to do something, I just text it with WhatsApp. I said, you know, whatever. R2-D2.

You know, there's this thing with the PowerPoint slide, slide number 17. That image is scaled or titled wrong. It should not say CX9. It should say CX10. R2D2 opens up PowerPoint, modifies it, puts it in PDF, sends it to me. Can you imagine that? Easy.

That's a future where your AI is not just a laptop, not just a tool, but your laptop becomes an AI, becomes an assistant all day long. Now, you don't want to necessarily run everything in the cloud because if you can run it locally, it's free. Just like laptops, just like phones, just like in your house. You know, why rent the television? You're going to use it every day. Why rent the washer dryer? You're going to use it every, well, hopefully once a week, okay? You know, why rent your refrigerator? You're going to use it every day. Why rent your assistant computer? You're going to use it every day. Just going to go out on a limb here and say, Jensen, well, I'm probably not somebody who does his own laundry. He has no idea how often laundry is done.

This man, I don't even know if Jensen has children. If you have children, that thing is running every day. That's just how that goes. Yeah. But this, like, what a perfect encapsulation, I think, of everything happening right now. Like, that's why the RTX Spark chip exists, is to make devices that can do that kind of AI stuff locally. It's what a lot of the Windows stuff is after. It's what Microsoft thinks is the future. Like, that is as good an encapsulation of a certain bet about the future of computing as I have heard anyone make, and I'm desperate to know what you think about it. Okay, so first of all, I have to just call out Hayden Field, who wrote about Build with Tom Warren. And she wrote this line that I think just contextualizes all of this.

especially when you think about microsoft and opening up breaking up and microsoft having to like be at some company here's what hayden wrote this year's build had the vibe of a freshly single divorcee posting a thirst trap on instagram and i know for a fact that inside microsoft there's furious debate about whether that's good or bad um so oh interesting Hayden will tell you that she thinks it's good, right? Microsoft is feeling itself. They're out of the relationship. They're just like, look at us, man. I've been in the gym. I've been eating clean. I'm bad, baby. Talk to Tom. Actually, Mustafa's going to be on the coder. And they're training their own models. They want to be their own company with their own ideas about the future of computing. The fascinating thing about Jensen talking about editing PowerPoint slides on his laptop. One, very much implies your laptop is open and available to do tasks on call for you, which is not the state of laptops when you're not around them. Just a weird example there. Two, I don't know if he's aware of this, but most people rent Microsoft Office.

Right? Like if the intelligence is built into your laptop, but it's using PowerPoint, you're paying that SAS fee for PowerPoint anyway. So even the example breaks down because the first thing Microsoft is going to do is bundle some intelligence into what they now call M365, which is horrible. And then they're like, let the PowerPoint AI do it for you, not the assistant on your laptop. So something about where the computer lives is very much like at stake, but also totally confused. I can't believe in asking where the computer should live since Alexa came out. Yeah. Yeah. And what does that computer look like? And where does it live?

a 15-year-old question at this point. And you shouldn't rent the intelligence because it can just happen locally in your laptop and that's free is like, whew, that's a lot, right? Because it's not actually free. You need a new RTX Spark computer that can do local inference at a fast enough rate. Which is going to cost many thousands of dollars. Yeah. Jensen, have you seen what you are personally doing to rent prices right now? So there's just a lot in there about where does the computer live? What applications does it have access to? How much those applications cost? Is it on all the time? And if it's an always-on computer with permanent access to your applications that you're already paying SaaS fees for, shouldn't that just live in the cloud anyway?

That all just feels like, oh, we made one big bet, but Microsoft is feeling itself, wants to talk about the future of computers. And so we're going to talk about this thing that exists too. Even though, isn't that chip like already out? Wasn't Sean on the Vertecast saying that chip is basically already out? Yeah, it was like a year ago, I think, when NVIDIA shipped the Spark AI supercomputer. Do you remember this thing? It was like the little gold Mac mini looking thing that they were like, this is a supercomputer on everyone's desk and it cost $4,000 and I forgot about it until this happened. Yeah, that's what happened.

Yeah, I think that Microsoft flexing that it has a bunch of its own ideas and it is no longer dependent on OpenAI. And by the way, here's Jensen Wong, you know, the most powerful man in AI with all kinds of ties to OpenAI on our stage, talking about not renting compute from our partner that will only rent you compute that we're breaking up with. Like, there's just some great Microsoft in there. Like, Microsoft at this moment in time. They showed off devices that don't run Windows. They actually run Android. Like, this project's a lot of stuff. And that's all agentic. All of that compute happens in the cloud.

Right? Yeah. Like, yeah. Where, if you're using, it's like a little credit card thing, right? Like a badge size. Yeah. And you just talk to it and it's a little agent and like goes and does stuff in the cloud. I continue to think that rules. I talked about this with Tom too, but like that idea of like the little employee badge that becomes an intelligent AI assistant. Like if you're a Microsoft, that is a perfect idea. Like I really, I earnestly believe that is, I don't know if any that'll come true or be real or ever actually work successfully, but it is the most like complete and most Microsofty thought about AI that I have seen yet. Right. And it absolutely presumes that the computer is in the cloud. Yeah.

Right? You don't need any sort of local compute if it's just a little badge with a display that is telling you what your agents are doing off in the cloud. And that is 100% opposed to why rent your refrigerator when I can just text my laptop, which is the open claw announcement. And Microsoft is basically using open claw. Yeah. So they're pointed in two diametrically opposed directions. I'm sure you can write some coherent thesis about like, you got to make every bet and like, some people are going to want X and some people are going to want Y. But Jensen being like, why rent this stuff? It's like, everybody wants you to rent the actual inference. Yeah, I don't think Microsoft is the only one confused about this for whatever it's worth. Like, this is very much what...

OpenAI is doing with OpenCloud. This is why OpenCloud is now part of OpenAI, right? Like this idea that we should have some piece of it that lives in the cloud, some piece of it that can use your computer, not just live on your computer, but use your computer, right? Like that's the turn is how do we get access to all of your apps? We just use your computer for you. And what should run locally? What should run on your own sort of private data? And then what should run in like a public cloud somewhere else?

No one knows, right? Like Apple is trying to figure this out with Siri and Apple intelligence. They have this whole private cloud thing that kind of works. And then they just punt things to chat GPT. And I think we're about to get new ideas about where all of that goes. No one has any answers to this. So I don't think Microsoft is like unique and being confused here. But it just makes me think we've been having this debate for a long time. And I've been asking people this question for a really long time where one of two futures is coming. Either everything is going to be a computer, right? Smartphone chips became so cheap and so readily available that what if we just put one in everything became the thing? Or you're going to have this one sort of centralized computer that you log into from everywhere, which is like.

sort of the Chromebook bet or is like what Alexa is doing, right? Alexa exists outside of any given device. You just have a bunch of devices through which you interact with Alexa. Everybody's trying to do both of those things simultaneously right now and it's just very confusing. The answer is going to be some sort of hybrid model. I mean, this is true of smartphones too, right? And I think it probably should be. Like if you have a smartphone with a slow processor, it doesn't matter how much compute is available to you in the cloud. You're not having the most fun, right? For sure. And so this has been going back and forth for years. I just think in this case...

as we've been talking about, like all of your data is in the cloud. All of the applications that the AI needs to use more often than not in the cloud, right? And it just seemed the web exists, right? Like they often need to go search the web for stuff. It doesn't seem like having the big local...

Even OpenClaw started off as CloudBot because it was just a thing on your computer that was calling Cloud a lot to use your computer for you. And people are just burning through their token allocations using these harnesses. So at some point, either NVIDIA has made a chip that is so good, it can run a good enough model locally and your laptop isn't red hot and your power bills aren't skyrocketing. Or the cost of inference will go down, which is whatever we'll predict over time. And this will be a moot point. You just pick whichever one works for you and that's fine. Yeah. I think, to me, the big question there is...

There is a next thing implicit in all of this stuff that Jensen is talking about and in all of the bets that even these PC makers are making, that if we can make the chips fast enough, then it's going to completely change the way you use all of these devices and not just what you use them for, but how you interact with them. One of the things Jensen said in this is that he wants to talk to his computer. Everybody has this idea that typing is now outdated, that actually you should just be yelling at your computer all the time. And if you believe all of this, there is going to come a time when the whole size and shape and form of your PC is going to need to change because of AI. To me, there's no evidence of that at all. Can you get a sufficiently fast chip to be able to do more things locally? I think it's a really fun and interesting question. I've been using all of these dictation tools, the whisper flows and spoken leads of the world as a typing replacement. And in many ways, they're really interesting. They're also irritatingly slow because they're all doing stuff in the cloud. And you download the local ones and they're enormous and still kind of slow because I'm on a Mac mini, which they work fine. This stuff is good and fine. But in order for the sort of set of things you can do to expand, you do need these faster chips.

But the idea that that is going to immediately turn into some brand new kind of relationship with our laptop, I think is just dead wrong. I just don't see it. All right. I'm just going to, here's, I have one counter example for you. Okay. Just a very simple point to make. The hottest consumer electronics product of the past six months is the MacBook Neo. Yeah. A very inexpensive laptop that runs an older iPhone chip that does not have any pretensions to AI at all.

And yes, I know people are going to tell me that people are buying Mac minis all around San Francisco to run up. Whatever, man. The hottest consumer electronics product in the last six months is the MacBook Neo. No question. Everyone is super happy with that thing because it doesn't have any of these pretensions about new form factors. It's just a really good laptop for not a lot of money that is absolutely fast enough. Dell just put out a new XPS at a cheap price to compete with Neo. And I believe it was Dell who said to us, we have to stop marketing these things as AI PCs because consumers don't want them. Yeah. Yeah. Very much so. So I hear you. Like, where does the compute live? Where does the logic live? Is a live question. I think it's been a live question since Alexa.

Right. We're going to put Alexa in the smoke detector was a real moment that we all lived through. And it was like, why? Why is why are there microphones all over my house? I need one in every room and you need to tell me where the computer is. Where does HomeKit live? An open question. I don't think anyone knows the answer to, including the people who make HomeKit at Apple. Yeah. But does it live on your Apple TV? Does it live on your iPhone? Who knows? Restart them both. It's like a real thing that happens all the time. It's like, where should the logic be? How diffuse should the logic be? Is the logic on someone else's server or is it on 15 different devices in your home? By the way, this is the argument to put Home Assistant on a Raspberry Pi. That's another episode of the show that we're certainly going to have to do at some point.

But I don't think any of these approaches to computing, even if they change where the computer lives, where the logic lives, change the form factor. I was on a flight to Chicago yesterday, and I played games on my phone. there's nothing about any of these agentic tools that can replace that. And if you can't replace that, then you still have a phone and then you still have a device that is capable of doing all the things and nothing about that form factor is different. And I haven't seen anyone gesture at that problem, right? It's like, oh my God, you can just talk to everything now? The form factor should be different. It's like, bro, I played whatever horrible golf game for two hours on the flight yesterday. Like, do you have that? Like, no, okay, I'm done with this. Yeah. No, I think one of my favorite things about this whole moment is I think there is a really strong case to be made that laptops are a perfectly solved problem.

Like we nailed it. Laptops are done. Stop monkeying with the form factor. We can do it. We can have all kinds of new ideas about lots of things. No, no. Keyboard in the front, man. Shout out Paul Miller. Keyboard in the front. What a true nightmare that was for 10 minutes. Was that an Acer computer that they did that? That they switched them around? There was a lot of them. It was like, it was one of our weirdest recurring segments on the show was keyboard in the front. Yeah, that was really something. But even like just reading back over this thing Jensen is describing where he's like, if I need my laptop to do something, I just text it with WhatsApp. I say R2D2, there's this thing with the PowerPoint slide, slide number 17, that image is scaled or titled wrong. First of all, to your point, this is pure business brain. Nobody has any ideas about what this would do for you other than

PowerPoint. Like, A, do this on your phone. Like, there's a real case to be made that it would be easier to make PowerPoint good on your phone to make this change than it would to do all of the things so that I can send this very long message to WhatsApp that a bunch of things will then go happen. I don't know if you've ever opened PowerPoint or Google Slides on a phone. That is a fast way to make your phone overheat and shut down. That is true. Maybe that is an impossible technology problem. But your point is correct. Like, this is only compelling if you're the managing director at an investment bank. Yes. And in no way...

does this require a different kind of laptop, right? I'm sure you've seen all of these memes of like, now everybody like leaves their laptop slightly cracked outside of the bathroom because their agent is running. And it's like, first of all, buddy, Google caffeine app Mac and you will have solved all of your problems. There are ways to make sure your computer doesn't turn off when you close the lid. Like, I don't know if you guys know this, blow your mind. But also it works on your laptop. It's fine. This is not, we don't need new ideas about laptops here. And I think this goes back to your idea of like, isn't the actual future of this that all of this stuff lives in the cloud? But even if it's running locally.

My MacBook Air running an M4 chip could do everything described there perfectly fine, except that Apple is bad at AI. Like, we don't need new gadgets to solve this problem. And I'm like, I'm so struck by we have phones and we have laptops. And actually, that might be the entirety of what the AI revolution requires. I mean, there's a lot of startups that want that differently. I'm just coming back to, what did I play? I played like a game where you connect a bunch of lines. It was called like Flow. It's like, it's just burned time while the baby was asleep on the plane. Like whatever, man. Can your weird pendant that wants to be my girlfriend do that for me? Because if it can't, like I'm still going to have a phone and I guarantee you my phone is going to have a faster processor, a better cell modem, a bigger battery than your girlfriend pendant. Yeah, I'm sorry. Like, and I have a wife. So phone wife, nailed it. Sorry, friend. I think the idea that what if we can make much faster laptops?

is super exciting. I actually think like the whole RTX part thing is gonna be really interesting. If they have made the thing as sufficiently powerful as they say and can bring the sort of Windows ecosystem up to the Apple Silicon level, that's cool and exciting and will enable lots of new things for people. Nothing in that to me suggests we need an entirely new take on what computers are. I just don't, I don't see it. Not to be a total conspiracist, this industry is desperate to get away from the Apple techs. And the way you do that is you get people to buy different hardware, running a different OS with different toll rates in the app stores, like whatever that thing is. And to do that, you need a new user interface paradigm because otherwise you're competing with the iPhone.

This is why Meta did the metaverse for whatever reason. They did the metaverse. But underneath it all, if you go back and read all that coverage, Nick here actually wrote a long piece at Pixel Envy about the metaverse promises. And he points out, Mark Zuckerberg is constantly talking about the Apple tax on Apple and how much he hates Apple. That's the whole impetus behind that whole disaster. Their bet was the next thing. They had no conviction about what it was. They just needed the next thing. They just wanted another platform that they didn't have to pay the Apple tax on and didn't have to care about Apple's privacy policies. And so they were like, that's it. We're helmets on your head.

It's definitely going to be it. And now, what if you talk to it? And so Meta has thrown away the elements on your head. Put out the glasses. You can just talk to that. They're going to build personal superintelligence. Everyone else is chasing pendants and wearables and rings and bracelets. Because if you can just talk to the computer, you don't need all the apps on the phone. You don't need that distribution. You don't have to deal with Apple and the Apple techs. And Apple could just end this overnight by changing how the app store worked. Right? And they would cement the dominance of iOS for another generation. But they can't because that's all their money. And so now we have the site. And maybe that's great. Right? Maybe this is actually the competition everyone needs. But you just see, like, oh, there's a weird dance here that has nothing to do with whether or not you actually need the new form factors.

right maybe they're exciting but you don't need them if you weren't trying to run away from the economics of ios in particular and by the way android has the same economics it's just google is like here's our assistant and no one can beat that right because siri has to date been a disaster there's a lot of like oh we can beat siri But very few people are rolling up into an Android phone and being like, we can beat Gemini or Android phone. Like that seems much harder than beating Siri. So all the emphasis on Siri and also iOS customers spend more money. There's a lot of reasons for this. We'll see what happens at WWC. But I look at this, all of this frenetic, let's reinvent computers and somewhere lurking in the background at different priority levels for every company is we have to get away from Apple. We have to get away from the Apple techs. Apple and Google both, but I think particularly the Apple techs. And there is this ongoing sense that those things are only going to become more entrenched, right? And so it's like, if we're going to do it, we have to do it right now. Otherwise, Apple will figure out the next form factor that it needs. And we'll get to this in a minute. But there's been a lot of...

new reporting about Apple's vision plan for smart glasses and the vision pro and all this stuff. And it's like, Apple is also reportedly working on a pendant. Like Apple has these ideas too. And so there is this sense of like, we have to beat this paradigm first or else the ones that have already closed it will just close it again. And that's what Gemini is doing, right? Like the fact that Google caught up so fast with Gemini is like an existential threat to everybody else doing this because it's like, oh, well, Google, if Google just wins the next one, the way that it won the last one, we're all in trouble again. Yeah. The same duopoly just can do it again. This is why they're all doing browsers. This is why they're all doing computer use.

They need access to your data as directly as Google has access to your data and a staggering number of applications. Who knows, man? I just think like, I'll use smartwatch as an example. Right, the Apple Watch is a great product. People love them. Google just put up a Fitbit Air. I'm wearing this whoop band that tortures me daily about, if I could just tell the whoop band I have a baby and it could stop telling me that my sleep was interrupted, that would be great. And it just like won't believe me. Like I'm like, this is not going to stop for 18 years. One of my favorite things about the Fitbit Air is that the first night it was like, gosh, you woke up a lot. I was like, I have a 10 month old.

asshole and now it just says every morning it's like you did great even though the baby woke up six times i'm really proud of you it's like fully changed its tenor about my horrible sleep it's great maybe i just don't want to tell the whoopin that much information about myself it'll be like you're less masculine because you care for your child i have a lot of feelings such a low t move to care for your child it's constantly telling me it's like no one wants me to 10x my testosterone like zero people in the world think that is a good idea for me in particular um anyhow uh the that set of wearables works There's competition in the market for screenless health trackers, and they all kind of work with iOS in whatever way they work. And if you want to display one notification and make something smarter than that, Apple stops you. And there's not actually great competition for the Apple Watch in the category of things that Apple Watch is. There's competition for things that are next to it. I'm a much harder core runner, so I'm going to go to Garmin or whatever it is. And there's competition for things that don't even pretend to do the things the Apple Watch does.

All these screenless trackers, the whoops and the fitbit errors, they're just doing what they can do inside the apps that they can run. And then they stop at Apple's garden walls. And I think no one wants to repeat that. There should be way more competition in smartwatches. And there kind of just isn't because every single phone maker has a smartwatch that works best with their phone. And that's pretty much where that ends. It's the same with headphones. Actually, this has happened to the headphone market in like the saddest ways. True. Not to talk about the headphone check here, but there used to be pretty vast competition in headphones. And now there's one set of wireless earbuds that work best with your phone. Yep. And they're often made by the manufacturer of your phone. And that is where that ends. And like, doesn't that suck? It does. And so like all these AI products, like I respect that there's furious competition to get to the next turn of user interface because that's usually how you break the form factor and you break the lock-in.

But boy, it seems like Apple's going to get there with Siri before any of these figure out how to do all the rest of the smartphone stuff. Yeah. And I also think you're making me realize that maybe that's the reason everyone is obsessed with reinventing your desktop computer from the inside out, because it's the only way you can get access, right? Like NVIDIA can't do the things it's describing on your phone because it can't. It violates the policies. Apple prevents that from being possible.

but you can get pretty deep system access to somebody's Mac and start doing stuff at that level, right? Like I've spent a lot of time with this app Raycast, which is kind of the same thing. It's like, it started as an application launcher and then it's like, okay, we can actually just take full system access and control all of your windows and control all of your settings. And all of a sudden it's like, oh, we've become a new way to use your computer. And you only get that kind of access on people's desktop computers. So if you need to change the paradigm, you have to change it in some sort of giant non-accessory way. You can't build dependent because Apple and Google won't let you. So you have to start with a desktop computer because it's the only way in.

Maybe that's the reason everybody's desperately trying to reinvent your laptop, because it's the only move they have. Yeah. I mean, this is why Microsoft is announcing the Surface Laptop Ultra with the RTX Spark, and they've got a whole lineup of their partners announcing it. They want to make the claim that doing local AI inference on a laptop is the future of computing, and the form factor can change slightly, but actually, we're going to start selling laptops with 108 gigs of RAM, right? And for the first time ever, there's a reason for that if you're not a heavy Photoshop user or whatever, right? Sure. And I just don't... First of all, 120 gigs of RAM in this economy? Come on.

I could buy a laptop ultra. That's where we are. But I don't know that just adding a ton of RAM and processing power to a laptop is going to make it a different thing than a laptop. Right. Right. Yeah, I'm all for more powerful laptops. I think they should probably keep being laptops. Keep it in the front.

Do it. Bring it back. Somebody, you cowards, if you believe in AI, just get rid of the keyboard entirely. What if it's just one giant-ass touchpad and a microphone? I mean, that's what Lenovo is for. This is why Lenovo exists. We're going to see the Lenovo concept of, like, a folding tablet with no input except a microphone, like, tomorrow. All right, we need to take a break before anybody gets ideas. Lenovo, that was a joke. And also, trademark, Nilay Patel, 2026. All right, let's take a break. Then we're going to come back. We've got Hypedesk. We have the lightning round. We'll be right back.

Support for the show comes from Satellite. That's S-A-T-E-L-L-A-I.com to get 35% off with promo code VergeCast. Satellite.com, promo code VergeCast.

We pulled all that data and looked at it and said that it would be the second half of 2027, automated traffic will become larger than human traffic. I was stunned when just a few months later in May, a team came to me and said, you won't believe it, but automated traffic is now past human traffic online. And that's just expanding like crazy. And if you extrapolate that out, five years from now, we think that... Automated traffic will be a thousand times human traffic online. This was a fascinating conversation, not just about the future of the web, but also about how one of the world's most important software companies is deciding which of its employees it can replace with AI. Listen to Decoder wherever you get your podcasts. This series is presented by EY. First Citizens Innovation Banking, formerly Silicon Valley Bank, is the bank of the innovation economy. The name may have changed, but for over 40 years, their commitment to bold ideas, relentless ambition, and delivering what's next is as strong as ever. So if you're wondering where the road to innovation starts, it's here. First Citizens Innovation Banking. Head to firstcitizens.com slash innovation to learn more.

All right, we're back. It's time now for the Hype Desk, where our friends Ross and Ashley come and tell us about cool things happening on the internet and in the world. Today, it's just Ross. Sad. It's okay. Ross, welcome back. Hello, hello. Ashley sends her regards. She's just flying back from a three-week trip in Japan, so excited to hear about that. That is cool. Wait, actually, can we rewind to last week? Because you were going to come on last week and tell us about 007 First Light, which is the only thing I care about in the world. Yeah, I'm actually surprised you are playing it as much as you are. And you're not playing it on PlayStation, right? You're playing it streaming.

So we talked about this last week when you weren't here and I had to have feelings about it only to Eli, who does not care about my feelings about James Bond games. But I don't have a PS5 or a current Xbox. I play all of my games on the Switch at this point, which says a lot about the kind of gamer that I am, basically. But this is the first game in forever that I was like, I will go buy a console just for this game. I love James Bond games. They raise the price of the console and then they put out a game they like.

Yeah, so it was $650, which is more money than I'm willing to spend on a James Bond game. Thank you to everybody, by the way, who sent me the link to the Woot sale that there was for like out-of-box refurbished PS5s for $300. They were sold out before I got there, but these are my people. What I ended up doing is spending $200 on a year subscription to GeForce Ultimate, which comes with 007 First Light, and I have been playing on my Mac Mini, and it rules. This game is amazing. Okay, just a lot there. I feel like the whole segment is how incredibly rolled by the tech industry you finally have been. Yes, yes. 100%. You're gaming on a Mac Mini with a subscription bundle from a service no one uses.

I don't know, man. I will defend NVIDIA GeForce. It is surprisingly good. It's great. Fair enough. My internet's very fast and it's wired. They're going to pull you on stage at WWC and be like, we found one. We found a Mac Gamer. Yeah. Well, ironically, you have to use like 11 other platforms in order to get this to work. I had to sign up for an account from the game developer. I had to get into my Steam account and I had to get a GeForce account. I don't know where I actually own this game, but I own it somewhere and it plays on my computer now. And it rules. And this game is spectacular. I'm literally barely through the tutorial of 007 First Light and I'm already having like a true blast playing this game.

It is one of the all-time great tutorials. And people I've talked to in the industry love it because it's so smart. Y'all heard me talk about some pre-record, but it is basically WarioWare for adults. It teaches you everything in this cinematic montage. The way you would watch in a movie, right? You see James in a basic training. He's punching, he's going to shoot, he's driving, he's getting better every rotation, but you see that rotation because they want to show the passage of time. They do exactly that, jumping from scene to scene in the basic training. This is like 30 minutes of the game. It's not a big spoiler, but it's one of the smartest ways to get people caught up on the basics of the game. Yeah. And the whole first act is basically a movie that you kind of participate in. It's like 80% cutscene.

and then you walk. It's like most of the game. And it works. It plays somehow for the first full hour of this game. But anyway, we're mostly not here to talk about 007 First Light. You have something else for us this week. What do you have for us? Yes. Well, I will talk about 007 all you want, but primarily, this is the greatest week in gaming. This is the one I'm really legitimately hyped and excited about. I got started in the industry covering E3. E3 is dead, but everyone kept showing up to Los Angeles anyway this week, and so they've kept it going as Summer Games Fest. So this is the week where every major company shows off basically all the games they want to release the next 12 to 24 months. Make huge hype circuits, all these trailers, hands-ons if you're out in LA. You can go to these little events as press and industry members.

It's great. It's exciting. We just started this week with PlayStation State of Play a few days ago. And then Friday is the big flagship recording, followed by an Xbox showcase on Sunday. And rumor has it, a Nintendo thing next week, which not confirmed, but every year they've done something around here. So it's not E3, but everybody's like, well, we already have the hotels booked, so we'll just keep doing stuff in LA. Honestly, basically. Like when E3 died, everyone's like, well, what do we do? We all go drink at the Figueroa, the DW Marriott downtown. We all have our friends here. People just kept coming and people just like, everyone started doing their own decentralized events. And now a surname Jeff Keighley, who runs the Game Awards, has his own live stream that's kind of like the flagship press conference of the week.

So what have we seen? Any super exciting games everybody's about? I feel like the only thing I hear about now is GTA 6. And the joke is like XYZ is going to happen before GTA 6. And GTA 6 keeps getting delayed. Did anybody announce anything that's exciting in the run-up to GTA 6? So GTA 6 is definitively in November. Everyone feels pretty confident about that. But as a knock-on effect, we're seeing games that were going to usually come out in that kind of October-November timeframe. They're all either getting bunched up in November. There were so many September announcements already. Or they're getting delayed all the way to like February 2027 just to get away from this. Like everyone is scared of being like crowded out by GTA. But yeah, so Sony State of Play.

It's a lot of first-party games. A few third parties are really exciting. The two that have come up more than anything, Marvel's Wolverine, which is an insomniac game from the makers of the Spider-Man series, and God of War Laufey, which is a next mainline chapter, but they don't want to call it a sequel per se. It's just a new God of War game that does not star Kratos. Is it still a God of War game? It is still a God of War game. Mild spoiler if you've not played the 2018 game. To put this in Neelai terms, by the way, this is like, we made a Madden game, it's about baseball. Yeah, that's what this feels like. It's just like you're going to watch The Penguin and you're going to like it. I'm not going to do that. Well, first off, The Penguin, very good, won awards. I know that everyone, you understand what I'm saying. Point taken, yes. This happens in parallel to the first God of War game. The first game starts with Kratos burying his wife. And in God of War Laofy, you play that wife. And she's woken up in the afterlife, which for a game that's been so steadfast in saying we're sticking to one mythology at a time, she's now in like the God afterlife, where like it makes us like Egyptian.

And we're assuming Japanese, probably not a Jesus. That might be a step too far, but like any kind of weird religious domination, these are apparently where the gods go to like kind of commingle. And she's very quickly captured. And there are two things that people cannot stop talking about. One, no surprise, not playing Kratos is a big deal. And two, very quickly, she just meets a gelatinous cube with a sword stuck in it. And that cube's name is Frank. What's the plan, boss? Not your boss, Frank. That's Jesus. It's very obvious what's happening here. It is gelatinous cube Jesus. But he goes by Frank in the world. It's a little weird the way it's supposed to. I love it.

I understand why it's divisive, but we saw 20 minutes of gameplay, and Frank, to me, was my favorite part. Just a gelatinous talking cube with a slight sense of humor that makes everything, it's already kind of weird, just that much more bizarre. I'm sort of in on that idea. Yeah, thank you, right? That gives me a little bit of, like, the Stephen Merchant thing in Portal that just, like, followed you around and made jokes at you. That's the best part of that game. I just want to hang out with this weird thing that just moves along above me as I play a game. Yeah, and when the first God of War remake came out in, like, 2018, it's the same studio that made that as making this, Kratos had a talking, severed head with him at all times that would just comment on the world and be sassy, and at first we're like, this is kind of weird.

I think it's going to work in a similar effect, but I do have faith in them to kind of pull it off, even if the way they kind of showed it off is weird. But long story short, I think the big theme of State of Play and PlayStation, everything they showed off was, we make really great single-player games. We make very cinematic games. And you will want to play this, as we all know. They're going to be exclusive on PlayStation. So you're going to want to buy that console and get a truly unparalleled cinematic experience in what you play. For $650. Yeah, I mean, there's that. The past, what, decade of games is like, you're going to play a live service game with 10 million other people and download DLC and show them your outfits. And I'm like, I don't...

I don't care about any of that, actually. I'm just going to keep playing Madden until you put out a great single player game. In this conversation, there's a lot of them. Yeah, there's a lot of it. I mean, the Wolverine game is going to be basically a standalone Wolverine movie that you have to get to play in. And if you've seen the Spider-Man games, it's very similar. It's a lot bloodier. I mean, you get drenched in blood after every combat. My one pet peeve is somehow his self-feeling factor like cleans up all his laundry as well. Like he will just be perfectly clean by the next scene. Tech can't go that far. But in the middle of combat, like they are going all in like Mortal Kombat levels of gore in the way he's fighting. That's right. Sony's going to put out an accessory that actually 4D splatters you with blood. You know, weird Sony would have done it. I don't know about modern Sony. Weird 80s Sony would have absolutely sprayed you with blood. A little robot dog that comes in and just like squirts you and just keeps track.

It's good stuff. Yeah, I agree. I think this idea of what if we just go back to games that are fun to play by yourself is an exciting and good idea. I think so. I love those kind of games. I will say, you will still get the DLC content as story expansions. You'll still get the skins and costumes. That never goes away at this point. But, Neil, to your point, it is just for you and anyone you take a screenshot of and share on social media. I only dress for myself, Ross. I always have and I always will. There's going to be a God of War DLC where you can play as the gelatinous cube, right? We can just put that on the calendar. I would. I'm sorry. It's Jesus. It's Jesus. The Jesus gel. That's the name of the DLC. Oh, we're all going to get so many phone calls after this.

All right. It's good stuff. Also, Ross, I don't remember if it was you or Ashley, but I owe the both of you kudos for recommending Widow's Bay. I was hoping you would say this. But don't say work. I haven't finished yet. I won't. But it is like, I cannot believe how good that show is in a way that like very few shows hold up the way that they have as long as they do. Like that show has no right to be as good as it is. And I cannot believe how much I'm enjoying it. So thank you for that recommendation. I love it. It is basically comedians doing the best honor or tribute to horror they ever could. Like everything they do, they try to do it in like the name of horror and the scary shit.

But they're comedians at heart. So there's like this rhythm and cadence that makes it just so much more watchable. Like it's a masterpiece in like suspense. Not a lot of jump scares, just a lot of buildup and dread. It really is. And I will not say anything. Please finish it, Neil. I'm so excited for you to get to there. Yeah, at some point we're just going to do like a full spoilers edition of Widow's Bay Hypedesk and I cannot wait for it. It's like three weeks until the finale. And yeah, we are watching the minute it comes out every day. Apple, Apple is on a roll. It is, we're not here to talk about that, but I am just so happy with everything I've seen on Apple TV. Totally. All right, Ross, good to see you, buddy. Thanks for coming. Cheers, y'all. All right, now let's get to the lightning round. Starting with our good buddy, Brendan. Oh, but first, costume change. Hold on.

It's time, once again, for America's favorite podcast within a podcast, the one on my shirt, Brendan Carr's a dummy. Available now in the Verge Shop. I cannot tell you the number of people who have earnestly requested that we put the Brandon Carr's a Dummy theme song on Spotify. We should put it on Spotify. I don't know how. I don't either. Bandcamp account? We can pull it off. Yeah. It's on our SoundCloud. There are no rules on SoundCloud. By the way, we still welcome new theme songs. Oh, absolutely. Even though we licensed to be able to goombas here. Or we bought it, I think. We welcome new ones because it is the most fun. Yes. The cinematic universe needs many soundtracks and theme songs. If we're going to put it on Spotify, we're going to make a playlist. Do you understand what I'm saying? 100%. He was so dumb this week, David. What did he do?

I'm just going to read you a sentence that he tweeted. And then we can get into it. But this fills me with an irrational amount of rage. Okay. Brendan this week tweeted. On X, the Everything app. You know the one that you use for all your credit card payments? Right, sure, yeah. Here's what Brendan said. Thanks to President Trump, millions of Americans are paying a lot less for internet. That's not true. That's the sentence. No, there's more. But just that's the first. He didn't actually even put a period in that sentence because he tweets like a teenager. You just know that's not true. In your heart, right? You know that's not true? No shock. No one's prices have gone down for anything.

So that's the first part. Okay, there's the rest of the tweet. The Trump FCC, which Brendan runs, quickly shut down a Biden-era plan that would have spiked internet prices. A new study now shows that consumers would have paid $5.6 billion more annually under the Biden FCC's plan. You will note that he does not explain what this plan is or what they shut down or where $5.6 billion come from.

None of this is true. I just want to be 100% clear. None of this is true. No one is paying less for internet access. In fact, if you just Google internet price increases, you find a long list of price increases that AT&T and Verizon have imposed over the past several years, including just like random regulatory fees that they think you should yell to government about. This is the nature of telecoms in America. They do not experience competition. They only increase prices. We pay higher prices for slower speeds than anywhere else in the developed world. This is just the nature of the thing. What is Brendan actually talking about? The Biden FCC made it so that you live in an apartment building. Your landlord could not bulk sign every tenant up for a single service provider.

It's called bulk billing, right? So if you're living in a big apartment building, this is true in New York for many years, you live in a big apartment building, Optimum would just buy your building and then you would only be able to have Optimum. That's not competition, right? Like you already see how this goes. Optimum would just go to all the landlords on your block and cut them a deal and they would take their kickback and all the tenants on your street would have Optimum. And Optimum is just the one that most people in New York have, but all the ISPs did this. Right. This is just nakedly a bad thing. I should not have to explain how that reduces competition, right? Like in people in dense urban areas, like getting the different internet service was a big deal, right? Like Optimum is not beloved in New York. When Fios came to New York City, which has happened in just a huge wave of corruption and lies, but Fios did come to New York City. This was like a revelation for tons of New Yorkers that one other company was trying to break in. And it's a long story, but...

moving to a building with Fios in like the early 2000s in New York was a whole thing because people wanted to get away from them so bad. But this is true across the country. So the Biden FCC says we're not going to do bulk building. The landlord cannot just tell every tenant what internet service they're going to have. The tenants have to be able to choose, and that means different internet services can come to your building and sign you up. And if they want to pay to lay the infrastructure, they can do it, but they can go get more customers. So Brendan undoes this rule because he has never met a regulation that would increase competition for his telecom masters that he did not hate, right? He is a monopolist at heart, and every time there's a regulation that would increase competition or even transparency, he gets rid of it. This is the man who got rid of the broadband nutrition label because he claimed it was too onerous for broadband companies to tell you how fast their internet would be and how much it costs.

He hates it. Or the thing where they had to tell you what fees they charge you. Like, paperwork is the enemy of branded car, I would say. Yeah, transparency, competition, he hates it. He just wants to lie to you that Americans are paying less for broadband than before. No one thinks, even the replies to his own Twitter, like, what are you talking about? On X, the Everything app, which should be a friendly space for him. All right, so what is this study? What is the thing that is making me furiously mad? What is this study that claims we're paying $5.6 billion less for broadband? Well, David, it was sponsored by the Bulk Broadband Alliance.

They hired a consulting firm called Cartesian and they put out a press release. A recent study from the Bulk Broadband Alliance found bulk arrangements provide lower cost, greater competition, and expanded connectivity. You don't say that the Bulk Broadband Alliance thinks bulk broadband arrangements are better than not, huh? The We Love Monopoly Corporation put out a study saying monopolies are great. One click to be like, what is this study? And it's a sponsored study from the trade group that wants to monopolize your broadband. Well, this is so dumb. So dumb. And he's the telecom regulator parroting the studies from the telecom industry.

He's the one who's like, no, this is a lie. Like, I'm actually going to look at what's happening here. The FCC will do its own studies, not the bulk broadband alliance, which doesn't exist, right? That is just the ISPs who started a trade group. There's no grassroots bulk broadband alliance.

The people of New York City are not like, you know what we should do? We should start a lobbying group to really push for our interests. Like, I hate having to choose an internet provider based on speeds and feeds. No, we're going to start the bulk broadband. You and me, David. Finally, consumer rights group that pushes for monopolistic broadband policies. Who will think of the landlords, Eli? Do you know what I mean? It is the dumbest thing I've... It just took me one second. Like, sometimes brand cars and army takes me a long time to assemble, you know? I got to weave an argument about free speech and censorship and, you know, what he's actually doing is test balloons to write... No, no, no. This one is a new study and then I Googled the study and it's the bulk broadband alliance sponsored the study.

A new study by the Make Internet More Expensive Lobby discovered that making internet more expensive is good. It's so transparently dumb and corrupt, right? Like, try harder, Brendan. Make me Google one more turn. Make me have to use an AI tool to discover what is actually happening, not the first Google result. Yeah.

Anyway, everyone in his own replies is like, what are you talking about? My internet's more expensive than it was. Every internet provider is constantly raising their prices. The speeds are not getting faster. Their service is not getting more reliable. My AT&T wireless is useless. I don't even know what we're doing in America anymore when it comes to connectivity. But Brendan is out here tweeting that collectively we paid $5.6 billion less than we were before under the Biden plan. If you feel that, you let me know. Because I certainly don't. Brendan, as always, you're welcome to come on the show and discuss why you think tweeting studies from the industry that you're supposed to regulate proves your point.

or really explain any of this at all, because I don't believe it. But as always, that's been Brendan Carr is a Dummy, America's favorite podcast in the podcast. Now with merch. Brendan Carr is a dummy. This is a totally random aside, but did you see, there was this New York Times story from like a week ago where they reviewed a bunch of cabinet footage to find how the Trump administration speaks to Trump specifically. The beginning of Brendan's tweet made me think of this. The Times went through, they said over a dozen hours of cabinet meeting footage to analyze how his administration spoke to him. On average, at least one of every six sentences either flattered Mr. Trump, gave him credit, or criticized his political opponents. One in six is just like, that to me was just so like illuminating about how this entire administration works that all anybody who works for Trump does is talk about how terrific Trump is because that's how you keep your job. Yep.

It's just what it is. It's funny. He's also just like running out of guys. You know, he's like the director of national intelligence. It's this guy. It's fine. Yeah. He'd say he was nice to me. As long as he's nice to me, he keeps his job. Anyway, I have a lightning round item for you. Both of my lightning round items for you are like loosely about Apple and loosely about WWDC. The big thing that's been happening.

over the last couple of months is everybody trying to sort of put the pieces together about what John Ternus, the incoming CEO of Apple, is going to do differently from Tim Cook, the outgoing CEO of Apple. So everybody's trying to read tea leaves about supply chains and new software ideas. And I would just point out that most of these decisions, especially for things in the supply chain, are going to way predate any CEO announcement. So like giving John Ternus credit for what's happening right now is probably not correct. But at any rate, one of the bits of news that came out this week from Ming-Chi Kuo, who is a person who's very...

tapped into supply chains uh gets a lot of things right gets a lot of things wrong supply chains are an imperfect source of information um but a bit that has come out is that basically uh the apple vision pro as it is sort of currently constituted maybe no longer and that actually apple is pivoting much more toward uh simpler lighter smart glasses things as opposed to like full-on headsets um mark german at bloomberg sort of disagrees with this but only lightly uh he said over the weekend that apple is developing let's see it was quote a slimmer and lighter headset to succeed the 3499 vision pro uh That can mean a lot of things. But at any rate, it seems like... Gruber disagrees with this, maybe more strongly. John Gruber, he believes that there will be another Vision Pro. Yeah, there's a lot in the wind here, but it does seem like Apple is very focused on the idea of smart glasses as a thing, right? Like, Meta was sort of the first to accidentally discover that there's a real market here for this thing, and now Apple is falling further into that. So I think...

To me, this is one of the things about Apple's entire AI plan, but also about WWDC that I'm very curious to see. Like if Apple is moving away from pass-through VR headsets as the future and toward smart glasses, that would just, in theory, predict a lot of UI change everywhere, right? Like Liquid Glass last year was based on the idea that what if everything was kind of like the Vision Pro? And now if we're pulling back to something different, something more... voice-based like apple's apple's dictation is trash apple's whole voice-based input system is really bad siri sucks so like if you believe in smart glasses there are a bunch of ai decisions you have to make in order to make smart glasses work and i'm just very curious to see how that stuff starts to pan out. But what do you make of all of this Vision Pro smart glasses noise right now? First, I would just like to point out that I was correct about the Vision Pro in my review. I might have given it too high a score, which I know you will point out. I gave it a 7 instead of a 6. But, right, this is the endpoint of head-mounted display VR was the thesis of that review. This is the best one of these you can ever make, and it's definitely not enough. And I don't think it was enough. I mean, I understand it's very expensive and maybe cost them down, but Meta's running away from this. There's just, there's an endpoint to VR.

This is the best one of these you can make. It's the best one anyone's ever made. It's hard to see how it can get better except for getting lighter. And nothing in here is as compelling as it should be to make everyone want to live in a helmet all the time. So I want to say I was right. I think Apple always wanted to make actual glasses. And carrying on down, we're going to occlude your vision with OLED screens and then pass through cameras. That is a lot of technology and a lot of development that just points you in the wrong direction. You will make a lot of decisions downstream of that that will help you zero if what you actually need to do is put content over real light passing through lenses.

If you can point a camera at something and then the image from a camera gets to pass through your compute and you can get to recognize what it is and layer your AR information on that and then display a final composited image, that is a very different pipeline and set of decisions and all of that is just radically different. You can do it with more latency. There's all kinds of stuff you can do there that the real light is going to pass through my glasses lenses and all of these decisions are going to happen in real time such that it's going to put a label on a Washington Monument or whatever it is. That's the goal. You're in the art museum, you're looking at the painting and it...

puts the information next to the painting and you're not going to capture it through camera, you're just looking at it through a pair of glasses. That's the goal. And if that's your goal, you're not going to spin your wheels on we're doing real-time compositing of camera signals. Well, I think Apple's bet seems to be, and we talked about this at the time, that if you build the VR bit of it first, that you can then turn a $3,500 technical achievement into a $600 pair of smart glasses because you've solved some of the tech problems and now all you have to do is make it cheaper over time. Whereas what Meta, I think, accidentally discovered is that there are actually...

for a lot of people, killer use cases in the shitty cheap thing, right? Like Apple's bet was we have to do like, we have to do the Tesla Roadster in order to get down to the Model S, but actually like Meta made a bicycle and it turned out people wanted that. And I think the thing that no one in this industry saw, including Meta, was that headphones and camera would be enough to make people spend a lot of money on a pair of smart glasses. And I don't think Apple saw that because I think if Apple saw that, they would have done smart glasses first. First of all, shout out to every product manager who listens to The Verchest who's thinking about the skateboard metaphor of how you build a product.

Yeah. You don't start with a truck, you start with a skateboard and you build your way up. Every PM I've ever talked to you talks about that. Meta did that on purpose, by the way. Boz, Andrew Bosworth, often talked about how they were building Oculus headsets and glasses, and they were going to find out if they could cut one down or build one up.

Yeah. Meta tried to do both bets. They did both bets very obviously and very openly. They were like, we're going to see which one gets there first, but we're starting at both ends and we're going to head towards the middle. And I think they discovered that they had more killer use cases with starting simple and trying to build up. And that's what they're doing. There's meta display glasses and all that stuff now, but the problems are still the problems, right? Like you still need to put a camera next to those lenses, like identify the real world, pull in a bunch of contextual information, composite something onto a display over the right thing positionally for your eyes. Like that stuff is still essentially impossible and they haven't really done it yet. And even the meta display glasses.

they're still basically just doing like notifications and like, you know, like fake AR where they put like the map arrows in front of you, but they're not pointing to anything. There's something there that I think Apple can get to very quickly and obviously Apple has richer access to notifications and they've got better camera ideas and all this stuff and they've got their own halo and privacy, but they haven't solved the AR piece. And I think if they walk away from the Vision Pro, which I don't think they will, I think they will announce Vision OS 27 or whatever it's going to be and we'll have some new features and they will keep seeing what's there, but I don't think we're going to get a Vision Pro 2. I think we will see another product that reuses those ideas and maybe we'll get one more chipmunk to the Vision Pro. Like, I think...

If you see the opportunity in glasses, in smarter AirPods, the problems are so different that continuing to try to iterate on the solutions as they exist in Vision Pro, even at Apple's scale, with Apple's resources, with Apple's talent, you would just reprioritize all of that towards solving the harder, more lucrative problem. Totally. Yeah. And the one that everyone seems to increasingly believe is the thing, right? Like, even if you just look at the sort of traditional Apple thing is to go into a market that a lot of people think is the thing, but doesn't yet have a great product and try to build that great product. That's where smart glasses are right now. Everyone in tech, at the very least, believes smart glasses are going to be a thing. I think.

my stance on that opinion is much less certain than I was about like MP3 players before the iPod. But that is a sort of widespread belief in the tech industry is that smart glasses are going to be a thing. But this is more we're chasing a form factor. We're just trying to get away from the iPhone. Sure. But again, I think Apple is also in a position of being like, actually, maybe the best place for smart glasses to be is an accessory to the iPhone. And I think Apple's probably right about that. It's just that no one else can do that. Like you said, Apple has access to the notifications. It can do more stuff with the camera. Like if anyone can build smart glasses.

to do all of the things you would want smart glasses to do, it probably is Apple. Yeah. Which is a bummer. I do feel like I should gesture at the idea that there's some coming set of European Union regulations that might make all this more interoperable. Yes. It's on the side. The Europeans who work at The Verge are always talking about this stuff at this point. But, you know, it's like, sure. All the batteries are going to be interchangeable. Sure. You know, like, sure. Yeah. So maybe this market will crack open more because there is some EU effort to...

even notifications like uh the google smart glasses that they showed me at io i was like is this gonna work better with your iphone camera like yeah it's just open now like apple just like made that more open um so there's something changing in that it's not enough and certainly apple will find the one thing it can definitely lock down that makes its product better than everyone else's that's apple's great at doing that um i just think I think consumers are going to look at everyone wearing these glasses and react in a really weird way. Actually, here's a good example. I have been telling the joke on this show about how glasses that show me names and faces are the killer app for me personally for 10 years. Not a joke. Real thing that you believe. True fact. Everyone should be mad at Eli for the dystopian hellscape he wants to bring in just so he can have smart glasses. But the reason I'm saying it's a joke is because the other half of it is I'm aware that it requires a dystopian hellscape to exist. Fair. Right? Like, I'm constantly pointing out this thing that would make me reconsider the dystopian hellscape. But the dystopian hellscape is the reason I keep bringing up the example.

I would love names and faces. I'm extraordinarily aware that you have to build a dystopian surveillance telescope to do it. And anyone who's listening to the show is probably aware of that because we're talking about it all the time. I mentioned this in passing to Joanna when she was on the show talking about her book and we just talked about it for a while. And then someone made a clip of that and it went viral. It broke containment out of our audience. And all of the comments were like, this dummy doesn't know what he's talking about. I would never. This tech bro. And it's like, yo, I've been talking about this for the longest time.

And I'm extremely aware of the surveillance hellscape. That's why I keep using the example because it's the one that personally for me is like, oh, that'd be great, but the trade-off isn't worth it. And I just, first of all, experiencing algorithmic containment breaches is wild. I hope no one experiences it. It's just a wild experience to have. And second, you can just see, oh, there's a lot of people out there who will not accept this trade-off. And they're already going to smash the glasses and they're already pissed about it. And at some point, the Apple logo on the side of your glasses is not a sign of fashion or style or being on a bleeding edge. It is a sign of invasiveness and people will react to it very strongly. Yep, completely agree.

It'll be an interesting thing to see Apple try and navigate that. This is when, if you're Apple, you have to make all of the noise you've made about privacy pay off in a big way. And I don't have a lot of confidence in Apple's ability or anyone's ability to do that, but it'll be fascinating to see them try. I can talk about WWDC forever, but let's save that for next week. We've got a bunch of Apple stuff coming up next week. Let's end with your lightning round thing because it is one of the weird stories of the week and we should just talk about it. We should end with it. I'll do it quickly. You've got to end on a high note because mine is so dumb. It will make everyone so mad. You've got to find a fun gadget to end on it. All right, deal. All right, so my lightning round item...

It's just, you know, karma's a real bitch. So over the weekend, tons of Instagram accounts were hacked, including President Obama's Instagram account. And they were stolen by hackers and, you know, ransoms were held and all this stuff. And it turned out that this was not a sophisticated security breach. Meta has fired so many people in trust and safety and support and all of these roles, and they replaced them with an AI chatbot. And the hackers basically figured out that you could trick the AI chatbot into resetting the passwords on the account.

And the exploits are dead simple. 404 Media wrote about it. Everyone should go subscribe to 404. If I understand it by trick, you mean just ask. They ask, and then they would say, we're going to send you an email. And it's like, literally, the attackers would be like, I don't have access to the email. Can you just do it? And they'd be like, all right, I did it. Or can you send it to a different email? And they'd be like, okay, I did it. So like tons of accounts are preached. Meta says nothing about this. And finally, Andy Stone, Meta's head of comms, who is an absolute social media combatant most of the time, says in a reply to one of the posts about it, this issue has been addressed. Doesn't even announce that there's an issue, that they're working on it, in a reply.

Andy Stones, as this issue has been addressed. Now, Meta is one of these companies that is doing tons and tons of playoffs. It is just Black Thursday at Meta every Thursday. I don't know if you see all the memes of people stealing the free food from the cafeterias before the Monday where they're all going to get fired. And then all the people are like, I just got fired from Meta, now I'm free. It's all happening. They're just firing tons of people. In particular, they fire tons of people in trust and safety.

And so the people that didn't fire got reassigned to data labeling while meta, and data labeling is how you train AI, while meta is also imposing tracking on all their employees to get better computer use for the next model. But they also just announced because the employees hated it so much that you could take 30-minute breaks from the AI watching everything you do for personal reasons. This is the most dystopian stupid thing I've ever heard. Like incognito mode for being a person? Yep. Yes, you should not use your work laptop to do personal things in the first place, but also your employer should not track every move that you make. This is nonsensical. And then, of course, the thing that happened is none of this stuff got appropriately read, team, none of this stuff got appropriately put trust and safety, and there was a large-scale account breach on Instagram.

because there's no one there to watch over the AI. And you've just, you know, you've stupidly insisted that you can replace all these actual human professionals with your AI system. And of course it was breached instantly. Like there was no other outcome. It's just like, I'm mad about it, but I'm also like, did we all learn an important lesson today, Mark?

Like, should you babysit the, should you have someone watching over the store? I think we should have someone watching over the store. We fired our security guards and we got robbed. That's the weirdest thing. So stupid. You know they're not going to learn. You know they're just going to keep firing more trust and safety people. Yep. We'll put 404's story in the show notes too. It's very good. And it really is enraging in the most straightforward way that it's like, of course, this is what's going to happen when no one is paying attention. And like, this is, you want the AI takeover? Here's what it is. This is what it looks like. Yeah. And it's actually particularly Mark Zuckerberg has turned his back on content moderation and trust and safety so significantly.

Instagram as a platform is just full of racism now. It just exists in a way that it didn't used to. They're degrading the experiences people have. They're degrading the security and stability of their service by just firing off un-coded features and gaping security holes. And they're firing other people that made that not happen because they're in the way. Something much worse will happen as they chase whatever next turn. Would you trust Mark Zuckerberg with personal superintelligence right now based on the quality of the experiences you have on Instagram?

No. This is your problem. Yeah, yeah. Password hygiene, people. It's very important. Two-factor authentication. It didn't work because the AI chatbot's like, oh, you forgot your 2FA code? Yeah, we're all pretty screwed. This is, yet again, more evidence for all your devices in the ocean. All right, I have some good news to end us on. And it's actually a question for you specifically and Becky, your wife, more specifically. Have you told Becky that Supernatural is coming back? I have not told her that Supernatural is coming back yet. Supernatural, which we thought was dead because of meta's complete negligence and lack of interest in VR and the metaverse, was going to just be left on the vine. It's actually being spun out as an independent company, Supernatural Health.

We don't have a ton of details, but it sounds like it's going to be run by some of the people who made it in the first place. Becky, at one point, was like a dedicated, quest-wearing Supernatural player. Supernatural is the VR fitness game. We should explain what this was. This was Beat Saber, but it was Peloton. Beat Saber with squats. Yeah, it was cool. I loved it. I played it in the pandemic. This got us through COVID in a very real way.

How enthusiastic about this is Becky going to be? I think she'll be into it. We still have a Quest 2 laying around that was our Supernatural machine and we bought the custom headband for it. The sweat-resistant faceband and headband. So maybe we'll dig that back out. Have you used a Quest recently? I have a Quest Pro sitting around. I put it on for the first time in ages. The things they have done to that OS are completely mind-boggling. It's so hard to navigate now. It makes no sense to me. They shuttered Horizon Worlds. The whole thrust of that product has changed. Obviously, the Quest Pro is discontinued. There's a Batman game that's pretty fun. Shout out to Bat to the Arkham. There's a lot of confusion in that world. I really hope Supernatural Health, a new company, takes advantage of being independent and how much people want the product.

to find a way to make their own dedicated hardware. That's the thing that will make me really excited. VR headsets are basically a commodity now. We're over a decade into the product as they should exist. Even Meta was able to make the Quest 3S cheap. I want to say it's cheap-ish. I realize we're in Ramageddon and all this stuff, but that's the thing that will make me excited. So they can get away from that platform, which has gotten deeply confusing. Obviously, they're going to have to support all their existing users already about Quest headsets, but I think they have a much brighter future as an independent.

I agree. I mean, there's also like it would be trivially easy in certain ways to put Supernatural on like the Switch, get sort of a Wii Fit thing going on with the Joy-Cons. Like it wouldn't be that hard to do. You can do it. You could do all kinds of wild stuff with like image recognition on people's webcams. Like there's so many cool things that Supernatural could do and be on these devices that doesn't actually require VR that I think might be awesome. But it might also be part of the fun is the immersion of it, I think, to some extent. So like seeing them try to sort out what is a 2D.

supernatural experience look like might actually be kind of fun. I'm excited about it. Well, I'm dying to talk to Chris Milk, who is the founder of Supernatural. I'm assuming he's part of this spin out. They haven't said any names. They haven't said who's running the company. So I think there's still some confusion there. But my understanding is he's involved in some way. You know, the reason Supernatural worked is because Chris Milk is like a famous music video director. And so he had all the contacts in the music industry to license the music. Yep.

No one else could pull it off because the music industry is generally so hostile to any ideas. But Chris was like an insider, so he pulled it off. So if he can figure that out, if he can figure out the music licensing again without Meta's money and all this stuff, Chris, you're invited on the show. Come on the show. Explain us how this is going to work because I think a lot of people are going to be really excited. Also, shout out V Song, who wrote the big story about how furious people were that Supernatural was shutting down and then did the follow-up interview with Lena Khan about how she basically was like, I told you so. She's very much giving herself credit for people spinning it out. As she said. I'm going to give her the credit too. But she has covered every inch of that story and she actually broke this news too.

Yeah, it's pretty awesome. I'm just excited it's coming back. I don't know if it's going to make me use my quest anymore, but just in the abstract, I'm glad it's coming back. All right, we should get out of here. Nilay, what's on Decoder this week? Mustafa Silliman, the CEO of Microsoft AI, is on Decoder. We're going to talk about his breakup with OpenAI. He's just spending some time with Tom after build. Spicy. Guys, he's excited to talk about what he's up to. And yeah, you get the sense they're very excited to just outwardly go to war with OpenAI after seemingly inwardly going to war with OpenAI. I'm also going to ask him if he thinks we're in the foothills of the singularity.

There's a lot to react to for Mustafa. I think that should become one of the decoder questions. How do you run your company? Are we at the foothills? Yeah, we've got lots more VertCast coming up next week. You're going to be at WWDC, right? I am. Okay, so I think the plan is we're going to do a live stream on the VertCast right after WWDC. You'll be running around in Cupertino doing briefings, meeting with people, so you probably won't be on that. But then we're going to follow up and do a whole hotline episode of lots of people's questions, and I think that's going to be you and me. We're going to do a bunch of two days later follow-up on all the WWDC stuff. There's always lots to talk about. People always have lots of questions. We also have another Apple thing coming on the show on Tuesday. It's going to be a very Apple-y week. This tends to happen. I'm going to ride back home from WWDC on that day that we're going to do that episode. So if you thought Loopy post-Ioni live was fun, you just get ready because he

Tired Nilay is a time on a podcast, and I'm very excited about it. It's going to be great. If you have needlessly complicated questions that Nilay is going to be unable to answer after a red-eye flight, call the hotline 866-VERGE-11. Send us an email, vergecastatverge.com. Again, do a lot of this on Monday. As you have questions during WWDC afterwards, if there's things you want us to talk about, feelings you want us to have, call the hotline. Send us emails. We're going to dig into as much of it as we can all through next week. Until then, the Vergecast is a Verge production and part of the Vox Media Podcast Network.

Today's show is produced by Eric Gomez, Brandon Kiefer, and Travis Larchuk. A reminder that to get all of our podcasts ad-free, plus all of our subscriber-exclusive newsletters, all of our other coverage, to get the whole Verge, the best thing you can do is subscribe to Verge.com. We will see you on Monday. We pulled all that data and looked at it and said that, you know, it'd be the second half of 2027, automated traffic will become larger than human traffic. I was stunned when just a few months later in May, a team came to me and said, you won't believe it, but automated traffic is now past human traffic online. And that's just expanding like crazy. And if you extrapolate that out, five years from now, we think that...

Automated traffic will be a thousand times human traffic online. This was a fascinating conversation, not just about the future of the web, but also about how one of the world's most important software companies is deciding which of its employees it can replace with AI. Listen to Decoder wherever you get your podcasts. This series is presented by EY. You work hard for every dollar. So put those dollars to work for you with SoFi, the everything app for smarter money. With SoFi Checking and Savings, you can easily track your spending and saving while earning nine times the national average savings rate. Get paid up to two days early with direct deposit with no account or overdraft fees and earn a cash bonus when you sign up with eligible direct deposit. Terms apply. Open your SoFi Checking and Savings account today at SoFi.com slash Vox Bank. SoFi Checking and Savings is offered through SoFi Bank N.A., member FDIC.

Support for this show comes from Wix. You can make a great-looking website with Wix, and you can do it your way. Whether you want AI to jump in or prefer to do things yourself, get a custom ready-to-use website in minutes with Wix's AI Website Builder, or choose from more than 2,000 templates. Get built-in solutions tailored to your business and enjoy easy, fuss-free domain registration, web hosting included. Wix powers more than 280 million businesses around the world. Because with Wix, you can own your individuality, create freely, and scale fearlessly. Ready to create your website? Go to Wix.com. That's Wix.com.